智能手机构成的生物医学成像系统的应用范围不断拓展，通过设计光学附件适配器，可以扩展系统的成像功能，从而进行医疗临床辅助诊断。智能手机成像系统已在皮肤科、眼科、口腔科及耳鼻喉科、妇科等有广泛应用，并取得良好效果。综述当前基于智能手机的医疗设备成像系统在临床辅助诊断领域的应用情况，分析其成像技术原理和特点，依据在不同科室的应用分析并总结了成像系统的特点，并对其未来发展趋势进行展望，旨在为后续研究和开发基于智能手机的医疗成像系统提供参考。

This work demonstrates a smartphone-based automated fluorescence analysis system (SAFAS) for point-of-care testing (POCT) of Hg(II). This system consists of three modules. The smartphone module is used to provide an excitation light source, and to collect and analyze fluorescent images. The dark box module is applied to integrate the desired optical elements and offers a dark environment. The cost of the integrated dark box mainly includes the upper cover, box body, lower bottom, fixture and some optical elements which is about $109. The chip module is used for fluorescence sensing, which is composed of an upper plate, bottom plate and cloth-based chip. Due to the integration of multiple smartphone functions, the SAFAS eliminates the need for additional power sources, light sources and analysis systems. The dark box and upper and bottom plates are made by 3D printer. The cloth-based chip (about $0.005 for each chip) is fabricated using the wax screen-printing technique, with no need for expensive and complex fabrication equipments. To our knowledge, the cloth-based microfluidic fluorescence detection method combined with smartphone functions is first reported. By using optimal conditions, the designed system can realize the quantitative detection of Hg(II), which has a linear range of 0.001–100μgmL?1 and a detection limit of 0.5ngmL?1. Additionally, the SAFAS has been successfully applied for detecting Hg(II) in actual water samples, with recoveries of 100.1%–111%, RSDs of 3.88%–9.74%, and fast detection time of about 1 min. Obviously, the proposed SAFAS has the advantages of high sensitivity, wide dynamic range, acceptable reproducibility, good stability and low cost. Therefore, it is believed that the presented SAFAS has great potential to perform the POCT of Hg(II) in different water samples.This work demonstrates a smartphone-based automated fluorescence analysis system (SAFAS) for point-of-care testing (POCT) of Hg(II). This system consists of three modules. The smartphone module is used to provide an excitation light source, and to collect and analyze fluorescent images. The dark box module is applied to integrate the desired optical elements and offers a dark environment. The cost of the integrated dark box mainly includes the upper cover, box body, lower bottom, fixture and some optical elements which is about $109. The chip module is used for fluorescence sensing, which is composed of an upper plate, bottom plate and cloth-based chip. Due to the integration of multiple smartphone functions, the SAFAS eliminates the need for additional power sources, light sources and analysis systems. The dark box and upper and bottom plates are made by 3D printer. The cloth-based chip (about $0.005 for each chip) is fabricated using the wax screen-printing technique, with no need for expensive and complex fabrication equipments. To our knowledge, the cloth-based microfluidic fluorescence detection method combined with smartphone functions is first reported. By using optimal conditions, the designed system can realize the quantitative detection of Hg(II), which has a linear range of 0.001–100μgmL?1 and a detection limit of 0.5ngmL?1. Additionally, the SAFAS has been successfully applied for detecting Hg(II) in actual water samples, with recoveries of 100.1%–111%, RSDs of 3.88%–9.74%, and fast detection time of about 1 min. Obviously, the proposed SAFAS has the advantages of high sensitivity, wide dynamic range, acceptable reproducibility, good stability and low cost. Therefore, it is believed that the presented SAFAS has great potential to perform the POCT of Hg(II) in different water samples.

为了降低微流分析成本，该文提出了基于智能手机检测压电基片上目标微液滴方法，并开发出相应的应用软件，通过分析微液滴颜色信息和几何信息定位目标微液滴位置，结合分析时间获取微液滴输运速度。为验证提出方法的正确性，以5 μL黑色墨水溶液微液滴为研究对象，进行压电基片上微液滴识别和位置分析实验。结果表明，该文所提方法和开发的应用软件可识别和定位压电基片上微液滴，在功率为27, 5 dBm时，第2~10 s段内微液滴输运速度为0, 416 mm/s。

针对智能手机玻璃盖板缺陷检测方法存在检测柔性差、良率低、检测时间长等问题, 提出一种改进YOLOv3的智能手机玻璃盖板缺陷检测方法。在特征提取网络方面增加通道注意力机制以解决缺陷特征不明显的问题, 在特征检测网络方面增加了104×104维度大小的特征图以解决缺陷多尺度的问题, 最后对模型进行剪枝减少模型参数, 提高缺陷检测速度。从智能手机玻璃盖板生产现场获得涵盖崩边、坑点、脏污和划痕等4类缺陷的图片构建缺陷数据集, 对本文提出的方法和Faster R-CNN、YOLOv3、YOLOv4等算法进行对比实验和分析。实验结果表明, 本文提出方法的检测平均精度均值（mean Average Precision, mAP)为81.0%, 检测速度为43.1 fps。相比原始YOLOv3算法, 检测mAP提升了3%, 检测速度增加了6.7 fps, 相比于其他深度学习算法, 检测速度和检测精度均有所提升。所提方法满足智能手机玻璃盖板工业生产现场缺陷高精度、高效检测的需要。

为实现现场快速检测及测试结果的远距离共享, 探索了基于智能手机的压电基片上水体中铜离子智慧检测法。声表面波驱动待分析微样液, 比色传感器与微样液中铜离子发生比色反应。智能手机捕获比色图像, 并进行图像颜色空间转换、(L, a, b)(L,a,b值分别代表一种颜色的亮度、红绿色度和黄青色度)颜色分量信息分析及等分细分定位, 确定待分析样液中铜离子浓度, 自行开发了基于智能手机的铜离子检测应用软件,并应用于压电基片上微流分析。标准比色点实验结果表明, 在0~30 mg/dL内,拟合曲线的标准偏差为1.37, 相关系数为0.995。

为了解决现场指纹图像采集过程中指纹背景颜色复杂、导致对比度和采集效率较低的问题, 采用智能手机设计了便携式指纹图像采集装置。一方面将手机连接单片机, 电动切换不同波长光源, 获得多光谱图像; 另一方面自动采集不同光源下的指纹图像。为确保指纹图像质量满足后续识别需求, 采用多光谱融合及方向滤波技术对图像进行了增强处理。结果表明, 当光源数目为4时, 采集及处理时间低于1s, 满足现场采集需求; 装置能够有效采集指纹图像并融合增强。该装置具有低成本、自动化及便携性高的优点, 有望在刑侦领域取得广泛应用。